Machine Learning & Predictive–Prescriptive Analytics
Calibrated predictive models linked directly to downstream operational decisions, including rolling-horizon booking control and economic evaluation.
Applied AI & Data Research Scientist
Jinho Cha, Ph.D.
I develop data-driven decision systems that connect machine learning, statistical inference, stochastic modeling, and optimization.
My research spans predictive–prescriptive analytics, robust decision-making, AI-enabled operations, large-scale empirical modeling, and high-stakes R&D. I focus on turning complex analytical problems into validated, deployable decision frameworks.
Research
Research Focus
A method-centered portfolio built around decision value, uncertainty, and empirical validation.
Statistical Risk & Robust Optimization
Overdispersed count models, Negative Binomial risk, distributional ambiguity, tail protection, sequential control, and model-misspecification-aware decisions.
AI Decision-System Orchestration
Dynamic switching among heterogeneous analytical modules under quality, latency, reliability, and transition-cost tradeoffs.
Mechanism Design & Digital Coordination
Equilibrium analysis, decentralized resource allocation, fairness, convergence, resilience, and smart-contract-mediated coordination.
Simulation & Empirical Decision Analytics
Stochastic simulation, policy timing, ROI analysis, and empirical calibration using national health and financial datasets.
Statistical Foundations
Statistical inference for truncated distributions, count modeling, probability-based decision analysis, and quantitative model development.
Python
R
pandas
NumPy
scikit-learn
LightGBM
PyTorch
TensorFlow
XGBoost
Java
Jupyter
Selected Publications
Published Research
Only published work is shown here. Under-review manuscripts are intentionally omitted.
2026
Machine learning-driven overbooking optimization under dynamic cancellation risk: Evidence from hotel revenue management
Computers & Industrial Engineering, 222, 112341
Reservation-level machine learning and rolling-horizon optimization on 119,390 hotel bookings. Rolling predictive control increased mean realized profit by 33.22% relative to capacity-only control.
https://doi.org/10.1016/j.cie.2026.1123412026
Distributionally robust automation under overdispersed count risk: Ambiguity sets, tail protection, and oracle-relative policy gap
Journal of Industrial and Management Optimization, 22(8), 4039–4087
Distributionally robust sequential control under Negative Binomial count risk, with an empirical illustration using 437 monthly S&P 500 jump counts.
https://doi.org/10.3934/jimo.20261432026
Optimal automation under overdispersed discrete risk: thresholds and hysteresis in a Negative Binomial model
Journal of Industrial and Management Optimization, 22(5), 2503–2554
Dynamic Negative Binomial risk modeling with endogenous automation, threshold structure, regime dependence, and hysteresis.
https://doi.org/10.3934/jimo.20260922026
Mechanism design and equilibrium analysis of smart contract–mediated resource allocation
Journal of Industrial and Management Optimization, 22(2), 997–1033
Decentralized resource allocation with equilibrium, convergence, fairness–efficiency analysis, and shock-resilience guarantees.
https://doi.org/10.3934/jimo.20260372026
Adaptive switching optimization for heterogeneous decision modules under operational uncertainty
Journal of Industrial and Management Optimization, 22(8), 4088–4137
A stochastic dynamic optimization framework for switching among heterogeneous decision modules under quality, latency, reliability, and switching friction.
https://doi.org/10.3934/jimo.20261442025
Modeling ROI in chronic disease management: A simulation-based framework integrating patient adherence and policy timing
BMC Public Health, 25, 4270
A 10-year stochastic simulation framework calibrated with MEPS and NHANES data to evaluate intervention timing, adherence, cost, and return on investment.
https://doi.org/10.1186/s12889-025-25279-3Experience
Professional Experience
2025–Present
Faculty, Computer Programming
Computer Science Division, Gwinnett Technical College · Lawrenceville, GA
Conduct AI/data research while teaching programming and computational problem solving.
2021–2025
Senior Researcher
Korea Research Institute for Defense Technology Planning and Advancement (KRIT)
Led data-intensive R&D spanning analytics, simulation, intelligent systems, logistics, and technology decision support.
Concurrent
Lecturer, Artificial Intelligence
Sungkyunkwan University, Graduate School of Advanced Defense
Delivered graduate-level instruction in applied artificial intelligence during KRIT tenure.
2017–2021
Assistant Professor
Department of Mathematics, Korea Military Academy
Conducted quantitative research and taught probability, statistics, linear algebra, optimization, and related mathematical subjects.
Earlier Career
Military Operations Analysis / Operations Research
Republic of Korea Army Headquarters / Joint Chiefs of Staff
Applied simulation, operational data analysis, and quantitative decision methods to large-scale mission-critical planning.
Leadership
Funded R&D Leadership
Project Manager
National R&D program led from problem definition through validation and implementation.
Co-Principal Investigator
Intelligent coastal-surveillance R&D integrating sensing, analytics, and decision support.
Additional Programs
Leadership roles across simulation, training systems, personnel analytics, and technology programs.
Education
Academic Background
- Ph.D., Industrial Engineering — Clemson University
- M.S., Industrial Engineering — University of Florida
- M.S., Industrial Engineering — Texas A&M University
- B.S., Electronic Engineering — Korea Military Academy
Recognition
Selected Honors
- Presidential Commendation, Republic of Korea, 2025
- Ministerial Commendation, Ministry of National Defense, 2023
- Army Chief of Staff Commendation, 2019
Contact
Research collaboration and industry opportunities
I am interested in applied AI, machine learning, data science, decision intelligence, and research-scientist opportunities where rigorous quantitative methods can create measurable value.